### Apparently un-physical result from BC

Posted:

**Wed Nov 14, 2018 6:05 pm**Hi All,

I've measured the salinity of the effluent from a sand column but am having trouble making sense of the results. The linear pore water velocity is 1.78 cm/min and the column length is 25.2 cm. With zero dispersion, this would lead to a transit time of just over 14 minutes. In the observed data, salinity starts to rise at about 14 minutes and peaks at about 21 minutes, which makes sense. Qualitatively, the breakthrough curve is smooth.

When I model the output, however (with only dispersion as a free parameter), STANMOD suggests that effluent salinity should start to rise after about 2 minutes and the simulated curve is nowhere near to the data. A reasonably close fit between the observed data and model can only be achieved if the velocity is allowed to vary, and results in a value of 1.11 cm/min (D = 0.49). In either case I find this hard to believe - in the first instance, how can a fluid moving at just under 2 cm/min possibly traverse a column of 25.2 cm in only a couple of minutes, and in the second instance how can it cover the same distance in 14 minutes when it is only moving at about 1 cm/min?

More model details: inverse parameter estimation using deterministic equilibrium CDE. Pulse input of 41 seconds. Zero initial concentration, zero production. Position of breakthrough curve/initial value of output position 25.2 cm. Sample times are the halfway point between sample start and end times.

Any thoughts on this would be much appreciated!

Tom

I've measured the salinity of the effluent from a sand column but am having trouble making sense of the results. The linear pore water velocity is 1.78 cm/min and the column length is 25.2 cm. With zero dispersion, this would lead to a transit time of just over 14 minutes. In the observed data, salinity starts to rise at about 14 minutes and peaks at about 21 minutes, which makes sense. Qualitatively, the breakthrough curve is smooth.

When I model the output, however (with only dispersion as a free parameter), STANMOD suggests that effluent salinity should start to rise after about 2 minutes and the simulated curve is nowhere near to the data. A reasonably close fit between the observed data and model can only be achieved if the velocity is allowed to vary, and results in a value of 1.11 cm/min (D = 0.49). In either case I find this hard to believe - in the first instance, how can a fluid moving at just under 2 cm/min possibly traverse a column of 25.2 cm in only a couple of minutes, and in the second instance how can it cover the same distance in 14 minutes when it is only moving at about 1 cm/min?

More model details: inverse parameter estimation using deterministic equilibrium CDE. Pulse input of 41 seconds. Zero initial concentration, zero production. Position of breakthrough curve/initial value of output position 25.2 cm. Sample times are the halfway point between sample start and end times.

Any thoughts on this would be much appreciated!

Tom